Right now I am primarily focusing on the Korean laryngeal consonant system. I am interested in:
Descriptive analysis of the fortis stop in Korean (especially regarding its relationship to lenis).
Fortis stops have been lesser-studied compared to the other two stops in (Seoul) Korean. It's been noted that fortis stops can contain creaky voice (often but not always visualised through a pitch contour that contains jumps). My questions surrounding this include if this is at all important to the contrast system for either perception or production, and what can we do to measure such jumps and avoid throwing them automatically away? By using an F0 ratio (together with Paul Boersma and Silke Hamann), we're trying to see if these jumps can be systematically captured. This is still in the early stages of research.
Neural network analysis of unmarked stop contrasts.
With Silke Hamann; trying to look more into the phonology of the stop system, this time with a focus on lenis. Lenis seems to be unmarked in terms of laryngeal specification - can we capture this? How? We're trying to do this using BiPhon-Neural Networks. Further research will hopefully also involve cluster analysis.
Conflicting cue experiments
What happens when we adjust the cues to be mismatching? Does a short VOT always cue fortis, or can another cue override it?